Exercises - Using the Jurisprudence Library Verity Search Engine

All exercises that follow except #22 and #23 are based on the following fictitious inquiry:

Sir, Madam;

Not being very familiar with internet sites, I would like to know if there exists jurisprudence involving an Agent of the Commission giving erroneous information to an employment insurance claimant. This results in payments being made to which the claimant had no right.

Please respond to me at

Jim@ditto.com

I thank you in advance and wish you the best.

The entry page menu of the Jurisprudence Library offers users only one choice as far as search domains are concerned. If you choose to use the entry page menu, your search will be restricted to a sub-collection of the CUB folder (see grey box). Roughly 4% of all CUBs have been singled out by an independent authority as being significant and then set apart in a separate "search domain" for more efficient searching. To search the entire library, choose the "extended search" link below the Search Domains box now. You may wish to bookmark it at this time.

In order to more readily read the following instructions and carry out the exercises, you will need to open a second browser. From this second window, you may access the extended search page of the library at the following address (URL): /eng/library/searchxt.shtml Extended Search Of The Juriprudence Library

Before starting, note that although numbers are given for the expected number of results in these exercises, these may increase with time as new decisions are added to the system.

Select the CUBs directory by clicking in the square box beside the phrase "CUBs".

Under "Search Options":

increase the maximum number of items to retrieve to "unlimited";
select "50" results per page;
leave sorting option on "score"
leave word variations option on "English"

Explaining search options:

choosing the maximum number of results that come up -
- notice that even though you ask for only 50 results to appear in the results lists, a greater number of files may match your query and the total number of matches is nonetheless displayed at the top of each results page;
choosing the maximum number of results displayed on a given webpage -
- the files that are served up for your viewing will be limited to the number you select
ordering of those results -
- results are displayed based on the sorting option you requested (by score or by title).
- by selecting "score", decisions that have the highest weighting (as determined by the number of occurrences of your search terms and the length of the document, amongst other things) will be placed highest in the list - this is also referred to as relevancy ranking;
- operators that are not automatically relevancy ranked include <all>, <stem>, <word> and <any>
- when using operators that are not automatically relevancy ranked, they are ordered by "title"
- by selecting "title", your results will, with some exceptions, be in chronological order since most of the document titles are numbered.
searching for word variations (carried out by default) -
- for example, you may identify a search term such as quit and even if the judge used another verb tense (i.e. quitting), the decision will still be retrieved;
- however, you must identify in what language you wish for that to take place as the search engine groups English and French decisions into separate collections
- failure to do so may result in 0 results even though all decisions are available in the library - not all decisions have been translated (try searching for CUB 19000 with different word variation language options selected: "French", "English" and "Both Official Languages").

The Exercises

Please complete the questions in the order in which they are presented. They have been designed to progressively lead you to a better understanding of how to use the search operators and wildcards of the search engine.

In the search field identified by the phrase "Search terms: ", enter the following search terms exactly as you see them in the examples below (do not type in the numbers identifying the exercise number!) then carry out the instructions that follow each search term or phrase:

information

Select "Search".
You should get more than 13000 results. Notice the percentage relevancy figures beside the results presented. Open up CUB 44824 and compare the highlighted terms to your search criteria and notice in what parts of the document they are found. The fact that there are such a high number of instances of the search term as well as the fact that these words appear in key positions of the document such as the header increases the ranking relevancy assigned to this result.
information

Change the word variations option to "Both Official Languages".
Select "Search".
You now get more than 17000 results. Open both examples of CUB 30583 and look for the highlighted keyword. Why have the numbers of results increased?

Answer
erroneous information

Change the word variations back to "English".
Select "Search".
You should have more than 250 results. The numbers of results have now decreased. Why? Would the results be different if we had left the word variations in French?

Answer
inform

>Select "Search".
You should get in excess of 9100 results. Open CUB 3348 and look at all the highlighted keywords. Do they correspond exactly to the word you entered into the search term field?

Answer
"inform"

Select "Search".
You should get more than 1760 results. Open CUB 55864A and look at all the highlighted keywords. Do they correspond exactly to the word you entered into the search term field?

Answer
INFORM

>Select "Search".
Compare this result with the numbers arising from the search carried out in exercise 4). Open CUB 3348 and look for the highlighted keyword. Apparently uppercase versus lowercase makes no difference to the results. Let's look at this a little closer in the next exercise.
Inform

Select "Search".
You should have a couple of results. Open CUB 2443 and CUB 2231 and look for the highlighted keyword. Note that there are no examples of the search term with all letters in lowercase or indeed all letters in uppercase.

NOTE the following operators (Boolean terms). See "how it works? how it works" for details of the function of these operators.

<AND>

<OR>

<NOT>

What follows are exercises which will assist in understanding the purpose of these operators in an advanced query.
erroneous information<AND>misled

Now you have roughly 30 results. Why? Compare results with exercise #3. Open up CUB 7257 and compare the highlighted terms to your search criteria and against the results you observed in exercise #3.
erroneous information <AND> misled

You still have the same number of results as in exercise #8. Does spacing seem to have an impact? Would spacing make a difference between the words making up the search phrase "erroneous information"? What if we accidentally put a space in the middle of a word (ex. "erron eous information")?

Answer

NOTE the following wildcards and how to access them on the keyboard.

Question mark = ?

Star = *
erroneous information <and> misl*d

Now there are more results. Open up CUB 40657 using the "[highlight]" link and scroll through the text to "mislead". Note the impact of the asterix on the results of the search (compare with exercise #9).
(erroneous information <or> wrong information) <and> overpay*

Now there should be roughly 125 results. Open up CUB 6860 using the "[highlight]" link then scroll through the text and observe the highlighted search terms as you go. Note that although the search term "overpayment" is present as is "erroneous information", that "wrong information" is not. Is this expected? Now open CUB 44389 and repeat the process. Note that all search terms are highlighted. The search engine looks for “erroneous information” AND any word starting with the letters "overpay..." in the text. Once all CUBs have been thus checked, it will go back and look for "wrong information" AND overpay*. CUB 44389 contains all the keywords but need not have. Which ones must be present in all the results and which ones need not be present in any given results?

Answer
(erroneous information <or> wrong information) <and> overpay* <and> employment insurance act

Now there is less than half the number of results. Open up CUB 44389 using the "[highlight]" link then scroll through the text and observe the highlighted search terms as you go. How does increasing the number of search terms in combination with the <and> operator impact on your ability to narrow your search results?

Answer
(erroneous information <or> wrong information) <and> overpay* <and> employment insurance act <and> commission employee?

>We now have significantly less results. Do these cases relate to the internet enquiry we were given to start with? Take note of the occurrences of the search term returned for the commission employee? query. How did the use of the "?" impact our ability to gather results? Try the same exercise without the "?", compare the number of results and review the highlighted search terms in both instances.

Answer
(erroneous information <or> erroneous interpretation) <and> overpay*

You should have roughly 100 results. Note that issues can be expressed in different ways with the consequence that you get different results even though your new keyword means broadly the same thing. So, in this exercise we have roughly 20% less results despite them being equally as relevant as in exercise #11. Two of the cases that come up in the exercise #14 results list also come up in exercise #11. What query could you create that would bring these two cases up exclusive of all other results from the two exercises?

Answer
erroneous <not> erroneous information) <and> misled <and> overpayment

You should have roughly 30 results. Notice that none of the results will include the phrase erroneous information and we know from the previous exercises that this phrase is found in the decisions. Hence, the usefulness of the <not> operator in eliminating certain possibilities is made clear.

We know that there are many less Federal Court cases than CUBs and so there are fewer cases to choose from even if they are generally of a greater jurisprudential value than CUBs. Therefore, if you choose to seek out cases at the Federal Court level, it may initially be wiser to broaden the scope of a search to increase the chances of returning results. This can be accomplished by limiting the keywords or phrases to the issue level. An example follows.

De-select the "Cub" domain and select "Federal Court" instead.
erroneous information <or> erroneous interpretation

You should have approximately 7 results. That represents approximately 98% less results than when the same search is carried out in the CUB collection. Notice that some of the results begin with letter "A" (signifying Federal Court of Appeal) whilst another begin with the letter "T" (signifying Trial Division - Federal Court).

One way you can limit searches within the Federal Court domain would be to seek out only Trial Division cases. An example of how to accomplish this follows.
Title <STARTS> T

You should have roughly 65 results. Notice that for all Federal Court cases that appear in the results list, the titles of the decisions start with a "T".

Should you then wish to combine the two queries in exercises 16 and 17, this can be accomplished through the use of parentheses. Parentheses "()" indicate the order in which the directions are to be performed; information within parentheses is read first, then information outside parentheses is read next.
(Title <STARTS> T) <and> (erroneous information <or> erroneous interpretation)

You should have 1 result. Notice that only the search phrase "erroneous interpretations" comes up highlighted. Search values used in combination with field operators are not highlighted hence the "T" in the title field is not visible in red. Why doesn't "erroneous information" come up in the results?

Answer
(Title <STARTS> A) <and> (erroneous information <or> erroneous interpretation)

You should have approximately 4 results.
(Title <STARTS> A) <and> (title <ends>94) <and> (erroneous information <or> erroneous interpretation)

You should have 1 result.
NOTE that there is a thesaurus feature available. To have the search engine search on all catalogued synonyms of a given word that is included in the thesaurus, use the <thesaurus> operator before your keyword. Change "Maximum # of items to retrieve: " to unlimited. De-select the "Federal Court" domain and select "Cub" instead.
<thesaurus> information
Type in the keyword as identified above.
Identify the highlighted keywords in CUB 64613.
<thesaurus> misconduct
Identify the highlighted keywords in CUB 44340A. Unfortunately, the thesaurus does not contain all possible synonyms for a given word (seven of the eight synonyms given are not highlighted and therefore were not recognized as such by the search engine).
<thesaurus> misconduct<not>misconduct
Identify the highlighted keywords in CUB 63868. This query provides you with a method of quickly identifying those cases containing just the synonyms of the search term selected.

Having gone through these exercises, we hope that you would now be in a position to answer the fictitious inquiry if it were up to you to seek out the decisions that relate to it. Clearly though, much depends on the user's ability to describe what they are looking for. This involves coming up with keywords that can then be used in conjunction with the operators we have just explored.

There are an infinite number of ways of identifying the reference material you seek. If you are an appellant, one strategy involves identifying the issue(s) and possibly the sub-issues for which you are seeking redress. What follows are some sources of keywords to describe those issues. Bear in mind that by no means are these the only words that may be used to describe these topics. They are but a starting point and can be accessed from the tables of contents or dropdown menus of the following webpages:

Quick Reference Tool
Judicial Interpretations
View from the Courts
Index of Jurisprudence

Answers to Exercises

Answer to exercise #2: Exercise

The number of results has increased because the search term "information" is spelled the same way in both French and English. Therefore, French documents containing the search term are included.

Answer to exercise #3: Exercise

The number of results has decreased. There are fewer occurrences within the documents of the two search terms "erroneous information" together as a unit as opposed to the term "information" on its own. We have effectively fine tuned our search by identifying what kind of information we are interested in.

Answer to exercise #4: Exercise

No. Variations of the search term are clearly in evidence. See "Searching for word variations"

Answer to exercise #5: Exercise

Yes. Use of double quotations ensures that no variations of the word (suffixes) are included in the results list.

Answer to exercise #9: Exercise

Additional spacing makes no difference in the first two instances. If, however, a space is inadvertently inserted within a word, there will only be results generated if the same typographical error is made in the text of a decision.

Answer to exercise #11: Exercise

This result is expected. In this example the operator <or>, when used with the two search phrases within brackets, only requires that at least one phrase must be in the text of a decision. With CUB 44389, we see that all three phrases can appear together in a result but they need not in order for a result to be generated. The search phrase which must appear in every result is any suffix variation of the term "overpay". The phrases "erroneous information" and "wrong information" need not both appear in the results but at least one must be in the text since the operator <and> was used in conjunction with "overpay*".

Answer to exercise #12: Exercise

The number of results has decreased because there are fewer occurrences within the documents of the four search phrases together in one document. We have effectively fine tuned our search by limiting our results to cases that deal only with the more recent Employment Insurance Act as opposed to the Unemployment Insurance Act. Cases that originated and were heard before the coming into force of the Employment Insurance Act will not have that phrase in the text of the decision.

Answer to exercise #13: Exercise

Yes, the decisions presented in the results compare favourably with the requirements of the ficticious inquiry we started out with. The number of results has decreased because there are fewer occurrences within the documents of the five search phrases together in one document. Use of the "?" wildcard effectively means we are looking for the plural of the phrase "commission employee". If our aim is to have the most possible results on the topic outlined in the fictitious inquiry, it makes more sense to not limit ourselves to cases dealing with multiple employees. In this case, more useful results are obtained by simply allowing the default word variations feature to operate.

Answer to exercise #14: Exercise

(erroneous information <or> wrong information) <and> erroneous interpretation <and> overpay*

Answer to exercise #18: Exercise

At the time of writing of these exercises, no Federal Court Trial Division decision contains the search phrase "erroneous information".